INDEX

Explanations

phrases related to speculations or uncertainties

oai_token-act-pair · gpt-3.5-turbo Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res_fs3072-jb

Configuration

jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_3072

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

3,072

Data Type

torch.float32

Hook Point

blocks.8.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 guiActiveUn

-0.86

inction

-0.72

resso

-0.71

Export

-0.71

robe

-0.70

ascript

-0.69

ļéĨĴ

-0.68

Password

-0.68

MpServer

-0.67

emaker

-0.67

POSITIVE LOGITS

 yeah

1.15

uh

0.95

 maybe

0.94

yes

0.94

 sorry

0.85

sir

0.84

 dear

0.80

 technically

0.80

 please

0.79

yeah

0.79

Activations Density 0.610%

phrases related to speculations or uncertainties

No Comments

No Known Activations